[JIRA] (JENKINS-59652) [kubernetes plugin] Protect Jenkins slave pods from eviction

[jonathan.pig...@gmail.com (JIRA)]

Title: Message Title

Jonathan Pigrée created an issue   Jenkins / JENKINS-59652 [kubernetes plugin] Protect Jenkins slave pods from eviction Issue Type: Improvement Assignee: Carlos Sanchez Components: kubernetes-plugin Created: 2019-10-04 06:34 Environment: GKE cluster master and node pools version: 1.14 Cluster autoscaler activated Jenkins master LTS installed with official Helm chart (1.1.24) Kubernetes plugin: 1.19.0 Priority: Minor Reporter: Jonathan Pigrée I have a sporadic bug occuring on my Jenkins installation for months now: java.net.ProtocolException: Expected HTTP 101 response but was '500 Internal Server Error'java.net.ProtocolException: Expected HTTP 101 response but was '500 Internal Server Error' at okhttp3.internal.ws.RealWebSocket.checkResponse(RealWebSocket.java:229) at okhttp3.internal.ws.RealWebSocket$2.onResponse(RealWebSocket.java:196) at okhttp3.RealCall$AsyncCall.execute(RealCall.java:206) at okhttp3.internal.NamedRunnable.run(NamedRunnable.java:32) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748)io.fabric8.kubernetes.client.KubernetesClientException: error dialing backend: EOF   I believe it was already reported in these threads and I understand that this is caused by an HTTP 500 returned by the kubernetes API: https://issues.jenkins-ci.org/browse/JENKINS-39844 - https://stackoverflow.com/questions/50949718/kubernetes-gke-error-dialing-backend-eof-on-random-exec-command   However, after further investigation, I am sure now that the bug occurs only when the cluster autoscaler is on and more precisely when the autoscaler scales down while a Jenkins build is running. It maybe an edge case.   I tried to set the annotation cluster-autoscaler.kubernetes.io/safe-to-evict: "false" on my jenkins slave pods but it didn't protect them. So I am trying now to setup a PodDisruptionBudget for each of my slave pod to protect them from eviction.   However, when passing the PDB into the podTemplate yaml it is just totally ignored.   How can I protect my jenkins slave pods from eviction?  Add Comment   This message was sent by Atlassian Jira (v7.13.6#713006-sha1:cc4451f) -- You received this message because you are subscribed to the Google Groups "Jenkins Issues" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-issues+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-issues/JIRA.202342.1570170872000.866.1570170900354%40Atlassian.JIRA.